Fix model card metadata
Browse files
README.md
CHANGED
|
@@ -1,3 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# TinyE5-L6-384
|
| 2 |
|
| 3 |
**TinyE5-L6-384** is a compact **384-dimensional text embedding model** built from
|
|
@@ -74,7 +93,10 @@ where model size, CPU latency, and deployment efficiency matter.
|
|
| 74 |
|
| 75 |
|
| 76 |
## Docker Example
|
| 77 |
-
|
|
|
|
|
|
|
|
|
|
| 78 |
services:
|
| 79 |
embedding-server:
|
| 80 |
image: ghcr.io/huggingface/text-embeddings-inference:cpu-1.9
|
|
@@ -84,17 +106,65 @@ services:
|
|
| 84 |
- ./data:/data
|
| 85 |
command:
|
| 86 |
- --model-id
|
| 87 |
-
-
|
|
|
|
|
|
|
| 88 |
- --pooling
|
| 89 |
- mean
|
| 90 |
```
|
| 91 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 92 |
```
|
| 93 |
-
|
| 94 |
-
|
| 95 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 96 |
```
|
| 97 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 98 |
## Attribution
|
| 99 |
|
| 100 |
Developed by **[GrowBit Labs](https://growbitlabs.com)**.
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
library_name: transformers
|
| 6 |
+
pipeline_tag: feature-extraction
|
| 7 |
+
base_model: sentence-transformers/all-MiniLM-L6-v2
|
| 8 |
+
tags:
|
| 9 |
+
- sentence-transformers
|
| 10 |
+
- embeddings
|
| 11 |
+
- text-embeddings
|
| 12 |
+
- semantic-search
|
| 13 |
+
- information-retrieval
|
| 14 |
+
- onnx
|
| 15 |
+
- int8
|
| 16 |
+
- minilm
|
| 17 |
+
- e5
|
| 18 |
+
---
|
| 19 |
+
|
| 20 |
# TinyE5-L6-384
|
| 21 |
|
| 22 |
**TinyE5-L6-384** is a compact **384-dimensional text embedding model** built from
|
|
|
|
| 93 |
|
| 94 |
|
| 95 |
## Docker Example
|
| 96 |
+
|
| 97 |
+
### INT8 ONNX
|
| 98 |
+
|
| 99 |
+
```yaml
|
| 100 |
services:
|
| 101 |
embedding-server:
|
| 102 |
image: ghcr.io/huggingface/text-embeddings-inference:cpu-1.9
|
|
|
|
| 106 |
- ./data:/data
|
| 107 |
command:
|
| 108 |
- --model-id
|
| 109 |
+
- rizwan3d/tinye5
|
| 110 |
+
- --revision
|
| 111 |
+
- int8-onnx
|
| 112 |
- --pooling
|
| 113 |
- mean
|
| 114 |
```
|
| 115 |
+
|
| 116 |
+
### Available Variants
|
| 117 |
+
|
| 118 |
+
All variants are published under the same Hugging Face repository:
|
| 119 |
+
|
| 120 |
+
```text
|
| 121 |
+
rizwan3d/tinye5
|
| 122 |
```
|
| 123 |
+
|
| 124 |
+
| Variant | Revision | Model |
|
| 125 |
+
| ---------------- | ----------- | ------------------- |
|
| 126 |
+
| Safetensors FP32 | `main` | `model.safetensors` |
|
| 127 |
+
| ONNX FP32 | `fp32-onnx` | `onnx/model.onnx` |
|
| 128 |
+
| ONNX INT8 | `int8-onnx` | `onnx/model.onnx` |
|
| 129 |
+
|
| 130 |
+
### Safetensors
|
| 131 |
+
|
| 132 |
+
```yaml
|
| 133 |
+
command:
|
| 134 |
+
- --model-id
|
| 135 |
+
- rizwan3d/tinye5
|
| 136 |
+
- --revision
|
| 137 |
+
- main
|
| 138 |
+
- --pooling
|
| 139 |
+
- mean
|
| 140 |
+
```
|
| 141 |
+
|
| 142 |
+
### FP32 ONNX
|
| 143 |
+
|
| 144 |
+
```yaml
|
| 145 |
+
command:
|
| 146 |
+
- --model-id
|
| 147 |
+
- rizwan3d/tinye5
|
| 148 |
+
- --revision
|
| 149 |
+
- fp32-onnx
|
| 150 |
+
- --pooling
|
| 151 |
+
- mean
|
| 152 |
```
|
| 153 |
|
| 154 |
+
### INT8 ONNX
|
| 155 |
+
|
| 156 |
+
```yaml
|
| 157 |
+
command:
|
| 158 |
+
- --model-id
|
| 159 |
+
- rizwan3d/tinye5
|
| 160 |
+
- --revision
|
| 161 |
+
- int8-onnx
|
| 162 |
+
- --pooling
|
| 163 |
+
- mean
|
| 164 |
+
```
|
| 165 |
+
|
| 166 |
+
For CPU deployment, the **INT8 ONNX** revision is recommended when model size and inference throughput are the priority.
|
| 167 |
+
|
| 168 |
## Attribution
|
| 169 |
|
| 170 |
Developed by **[GrowBit Labs](https://growbitlabs.com)**.
|